Use of a priori information in the identification of global nonlinear models - A case study using a buck converter

This paper investigates some issues related to grey-box identification. In particular it is discussed how a priori information about the system static nonlinearity can be incorporated into the model. Also, it is shown that in some applications involving nonlinear systems, there must be a tradeoff be tween the accuracy of the estimated model static nonlinearity and the quali ty of forecasts. These concepts are applied in the identification of a real dc-dc buck converter operating in continuous mode. Improvement in model gl obal stability was achieved by using simple a priori information that takes into account the steady-state voltage relation of the converter. Models th at are valid over a wide operating range are included. Such models are comp act and can he estimated directly from data obtained from the converter.